基于扩展卡尔曼滤波和GSA的多变量灰色经济预测模型  

A Multivariate Grey Economic Prediction Model Based on Extended Kalman Filter and GSA

作　　者：高志熙 韩晓红[2] GAO Zhixi;HAN Xiaohong

机构地区：[1]山西清众科技股份有限公司,山西太原030024 [2]太原理工大学,山西太原030024

出　　处：《山西科技》2020年第2期135-138,共4页Shanxi Science and Technology

基　　金：山西省自然科学基金资助项目(项目编号:201801D121136)。

摘　　要：提出一种多变量灰色预测模型。该方法使用扩展卡尔曼滤波(EKF)实现灰色预测模型的参数估计和预测,同时在扩展卡尔曼滤波步骤中引入万有引力算法(GSA)识别未知噪声统计特性。为了验证模型的有效性,利用山西省2003-2012年的国民生产总值(GDP)及其相关影响指标进行预测,与使用最小二乘法进行模型参数估计的多变量灰色模型(MGM)相比,提出的模型的预测准确性更高。A multivariable grey prediction model was proposed. This method uses extended Kalman filter(EKF)to estimate and predict the parameters of the grey prediction model,and introduces the gravitational search algorithm(GSA)to identify the statistical characteristics of unknown noises in the extended Kalman filter step. In order to verify the effectiveness of the model,the GDP of Shanxi Province from 2003 to 2012 and its related impact indicators were used to predict the model. Compared with the multivariate grey model(MGM)which used the least square method to estimate the model parameters,the prediction accuracy of the proposed model was higher.

关 键 词：灰色预测 EKF GSA MGM 国民生产总值指数 

分 类 号：F127[经济管理—世界经济]